All:
I need a sanity check on a design I am wanting to do. We have a Cobol STC
running with a Assembler Operator Interface, single thread.
I understand I could redo in Assembler and multi-task, redo in C or C++ and
thread, but this isnt what I am looking for. Inside the STC we maintain
a
Hi John,
I do have the Product libraries in the APF,LPALST00,LINKLIST. This are VSAM
files which are not allocating though the catalog are shared.
On Thu, Aug 27, 2015 at 5:42 PM, John McKown john.archie.mck...@gmail.com
wrote:
On Thu, Aug 27, 2015 at 3:22 AM, Nathan Astle
Good point. A path name can be up to 1023 characters long, including all
directory names, file names, and separating slashes. So 1023; I stand
corrected. The program name (defined how? as the final node of the path
name?) would have to be somewhat less, but certainly in the 21st century might
On Thu, Aug 27, 2015 at 8:09 AM, John McKown john.archie.mck...@gmail.com
wrote:
Ah. not a RACF shop. I remember long ago in an ACF2 shop that I had a
problem allocating the ACF2 rules data set. What I had to do there was
totally remove the catalog entry and start up ACF2 in fail safe mode.
Charles Mills wrote:
Isn't it now (since around 1993 or so) 1024 for a USS program?
Including / excluding path length?
Compare SMF30EXN.
Indeed. After RTFM in my SMF bookie (Ok, v1.12 for now, I will check v1.13 and
v2.1 bookies later),
I see this interesting snippet:
quote
The field
On Thu, Aug 27, 2015 at 7:49 AM, Nathan Astle tcpipat...@gmail.com wrote:
Hi John,
The System is not yet fully up and it is failing at the Security product
So I take it that you can't logon to TSO or a UNIX shell at this point on
this system.
start up. I can see the Catalog services are
No, not irrelevant at all. The original question was to determine when a
certain program was executed. The problem is that there is no record in SMF of
every program executed. Jobstep program names are recorded in SMF. The certain
program might not be a jobstep program, but it might well show
Or are you saying that the Program name running in the task that used the
largest percentage of CPU time in this address space could only be a legacy
load module/PDSE program object, not a program residing in an HFS directory?
My belief is that the IBM developers know what a task is and would
Just to get the water flowing I'd make a new Alias in Test system QWST and
try alt rename or just repro from QWS.* to QWST.*
What are you using for ENQ propagation? GRS, MSX?
In a message dated 8/27/2015 7:49:38 A.M. Central Daylight Time,
tcpipat...@gmail.com writes:
The System is
Also, if you are not aware, there is an ACF2 list on CA Communities that might
be better able to help
go to support.ca.com and select the COMMUNITY menu option.
or
you could also join the ACF2 list on Yahoogroups
https://groups.yahoo.com/neo/groups/ACF2-L/info
Lizette
-Original
Information I would need at this point is
1) Did you contact CA ACF2 for assistance? They are very helpful and can
probably resolve this issue more quickly.
2) The output of the following IDCAMS lists
LISTC ENT('QWS.SEC.RULES') all
3) Provide the section of SYSLOG where your IKJ message
Stretching my memory, isn't there something about VSAM files can only be
cataloged in 1 catalog, unless they are SYS1.* datasets? I think there
is some information in the VVDS?
Regards
Andrew Rowley
On 27/08/2015 22:24, Nathan Astle wrote:
Hi John,
I do have the Product libraries in the
On Thu, Aug 27, 2015 at 7:36 AM, Andrew Rowley and...@blackhillsoftware.com
wrote:
Stretching my memory, isn't there something about VSAM files can only be
cataloged in 1 catalog, unless they are SYS1.* datasets? I think there is
some information in the VVDS?
This is correct. There are
SMF30_Highest_Task_CPU_Program has the program name from the task that
reported the largest CPU time percentage for the step (in 30-4, if I
understand the manual correctly).
Why would IBM make it only 8 characters long? Sigh.
Because eight characters is the longest a program name can be.
Isn't it now (since around 1993 or so) 1024 for a USS program?
Compare SMF30EXN.
Charles
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf
Of Bob Shannon
Sent: Thursday, August 27, 2015 5:31 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re:
Hi John,
The System is not yet fully up and it is failing at the Security product
start up. I can see the Catalog services are available.
Nathan
On Thu, Aug 27, 2015 at 6:15 PM, John McKown john.archie.mck...@gmail.com
wrote:
On Thu, Aug 27, 2015 at 7:36 AM, Andrew Rowley
The program name (defined how? as the final node of the path name?) would
have to be somewhat less, but certainly in the 21st century might be well
over 8.
Irrelevant to the original point. SMF30_Highest_Task_CPU_Program is for tasks,
not for processes. If you need an SMF field for
Scott Ford wrote:
We have a Cobol STC running with a Assembler Operator Interface, single
thread. I understand I could redo in Assembler and multi-task, redo in C or
C++ and thread, but this isnt what I am looking for. Inside the STC we
maintain a secondary index of data items. If the
Well, the book (z/OS DFSMS Managing Catalogs) has this to say:
The system applies the parameters in IGGCATxx, LOADxx, and SYSCATxx in
the following order:
1. Parmlib member IGGCATxx, if specified, takes the highest priority,
followed by...
2. Parmlib member LOADxx followed by...
3. SYSCATxx
In every shop that I've worked in, the pool of 3270 devices defined as
consoles were solely dedicated to that purpose. A separate pool of 3270
devices not defined as consoles were dedicated as terminals. So I would be
perfectly happy to have an CONSOLxx option to formally dedicate the
consoles as
John McKown wrote:
Damn, gotta stop typing on phone email at same time. Change ASID to ACID.
Careful fella, multi-tasking is reserved for ladies, not us lame males... ;-D
Groete / Greetings
Elardus Engelbrecht
--
For
This is not my forte, but I thought I would just throw this out.
Could you STC be written with a higher level Assembler stub routine that would
be monitoring for things like CANCEL . That routine would then signal a
daughter task to provide the function you want?
Lizette
-Original
In a quick test it seems that it is recorded as *PATHNAM when SMF30EXN is sshd.
I don't think that fits even a Talmudic definition of program name.
Charles
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf
Of Bob Shannon
Sent: Thursday,
On Thu, 27 Aug 2015 09:02:57 -0400, Scott Ford idfzos...@gmail.com wrote:
All:
I need a sanity check on a design I am wanting to do. We have a Cobol STC
running with a Assembler Operator Interface, single thread.
I understand I could redo in Assembler and multi-task, redo in C or C++ and
thread,
On Thu, Aug 27, 2015 at 10:45 AM, Walt Farrell walt.farr...@gmail.com
wrote:
On Thu, 27 Aug 2015 09:02:57 -0400, Scott Ford idfzos...@gmail.com
wrote:
All:
I need a sanity check on a design I am wanting to do. We have a Cobol STC
running with a Assembler Operator Interface, single thread.
On Thu, Aug 27, 2015 at 10:57 AM, John McKown john.archie.mck...@gmail.com
wrote:
Basically, if it is critical enough, Scott really needs something which is
ASID, in the data base sense. Personally, I ain't talented 'nuf to be
writing such. Perhaps the dual, alternating DSNs suggested by
On Thu, Aug 27, 2015 at 6:02 AM, Scott Ford idfzos...@gmail.com wrote:
All:
I need a sanity check on a design I am wanting to do. We have a Cobol STC
running with a Assembler Operator Interface, single thread.
I understand I could redo in Assembler and multi-task, redo in C or C++ and
You cannot do LISTCAT without some extra software support, but you can issue
MODIFY CATALOG commands from the console. I would focus on two. Issue these
commands on both a working system and on the problem system.
First verify that you indeed have two-level alias specified. That is not the
Walt,
A big Ty, I like the idea of trapping the abend after the C command is
issued. I have routine for the F so all I have to do is code for C and P ,
correct ?
Regards,
Scott
On Thursday, August 27, 2015, Walt Farrell walt.farr...@gmail.com wrote:
On Thu, 27 Aug 2015 09:02:57 -0400, Scott
Can anyone get to the IBM Support Portal?
http://www-947.ibm.com/support/entry/portal/support
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO
I can get to the screen.
What do you mean by DOWN? are you getting an HTTP 404 type error?
Lizette
-Original Message-
From: Staller, Allan allan.stal...@kbmg.com
Sent: Aug 27, 2015 7:50 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: IBM Support Portal Down?
Can anyone get to the IBM Support
Receiving The page isn't redirecting properly w/Firefox.
Just tried it w/Chrome and all is OK.
Restarted Firefox and All is OK...
Thanks for your help...
snip
I can get to the screen.
What do you mean by DOWN? are you getting an HTTP 404 type error?
Lizette
-Original Message-
Elardus,
I like that also..I like the RACF protection...especially on a STC for
Provisioning and Reconciliation.
Regards,
Scott
On Thursday, August 27, 2015, Sam Siegel s...@pscsi.net wrote:
On Thu, Aug 27, 2015 at 6:02 AM, Scott Ford idfzos...@gmail.com
javascript:; wrote:
All:
I
On Thu, Aug 27, 2015 at 10:03 AM, Lizette Koehler stars...@mindspring.com
wrote:
This is not my forte, but I thought I would just throw this out.
Could you STC be written with a higher level Assembler stub routine that
would be monitoring for things like CANCEL . That routine would then
On Thu, 27 Aug 2015 07:16:34 -0700, Charles Mills charl...@mcn.org wrote:
No, not irrelevant at all. The original question was to determine when a
certain program was executed. The problem is that there is no
record in SMF of every program executed. Jobstep program names are recorded in
SMF.
On Thu, 27 Aug 2015 10:16:10 -0500, John McKown john.archie.mck...@gmail.com
wrote:
On Thu, Aug 27, 2015 at 10:03 AM, Lizette Koehler stars...@mindspring.com
wrote:
This is not my forte, but I thought I would just throw this out.
Could you STC be written with a higher level Assembler stub
Thanks All, I am going to code up a prototype and test on our Sandbox, so i
can see the results of a 'C' command for example.
Scott
On Thu, Aug 27, 2015 at 12:02 PM, John McKown john.archie.mck...@gmail.com
wrote:
On Thu, Aug 27, 2015 at 10:57 AM, John McKown
john.archie.mck...@gmail.com
Now that your system is up, be sure to 'harden' the dynamic change. These days
alias level is specified in LOADxx in your IPL PARM data set. It defaults to 1
and may not be currently specified at all, which would explain your string of
woes.
.
.
.
J.O.Skip Robinson
Southern California Edison
Ah, we don't use IGGCATxx. There is also the ancient SYSCATLG member of
SYS1.NUCLEUS. I'm sure that LOADxx overrides SYSCATLG, but what about IGGCATxx?
If these members specify different values for ALIASLEVEL, who wins the thumb
wrestle?
.
.
.
J.O.Skip Robinson
Southern California Edison
Lizette,
That's a idea. I guess maybe CEEPIPI..I want to make sure I do this so I
don't cause other issues for customers, that would not be good.
Regards,
Scott
On Thursday, August 27, 2015, Lizette Koehler stars...@mindspring.com
wrote:
This is not my forte, but I thought I would just throw
W dniu 2015-08-19 o 00:26, Robert Harrison pisze:
From technologyreview.com:
http://www.technologyreview.com/news/540011/mainframe-computers-that-handle-our-most-sensitive-data-are-open-to-internet-attacks/
Really?
What I understod from the lecture:
a) mainframes are old, obsolete, but
jo.skip.robin...@sce.com (J O Skip Robinson) wrote:
Now that your system is up, be sure to 'harden' the dynamic change. These days
alias level is specified in LOADxx in your IPL PARM data set. It defaults to 1
and may not be currently specified at all, which would explain your string of
woes.
You can trap cancels with ESTAE(X) but interfacing back to LE is not for the
feint-hearted. Sx22's are unrecoverable (note the quotes -- you can do it,
but it's not for the feint-hearted).
Charles
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU]
Might be related to the number of devices defined to the environment.
Are you talking about above or below private?
Mark Jacobs
Phil mailto:philyo...@gmail.com
August 27, 2015 at 2:32 PM
Hello All,
What could cause private area to shrink 1M when the system is IPL'd on
a different machine as
I've seen this happen when adding a module to CSA and it causes CSA to cross a
1-meg boundary. CSA expands in 1meg increments.
Cliff McNeill
From: IBM Mainframe Discussion List IBM-MAIN@LISTSERV.UA.EDU on behalf of
Phil philyo...@gmail.com
Sent:
IIRC RMFPP has one but don't think it identifies who the owners are?
In a message dated 8/27/2015 3:00:13 P.M. Central Daylight Time,
stars...@mindspring.com writes:
I was looking for a graphic that displays this better than I can write it.
So private area below the line is what is left over after the system carves out
everything else.
So CSA, HSA (I think - Hardware area), etc
I was looking for a graphic that displays this better than I can write it.
So
The portion of the user private area in each virtual address space that
Hello All,
What could cause private area to shrink 1M when the system is IPL'd on a
different machine as in, in DR? Assume that the image being IPL'd is identical
to production. I'm told that there were no changes to parmlib, LPA, nucleus
etc. Thanks.
Have to examine all the parts. CSA, ECSA, etc that take away from private.
Something like Mark's IPLINFO is a starter or the PRIVAT mapper in IVP lib
is another choice.
In a message dated 8/27/2015 1:47:44 P.M. Central Daylight Time,
mark.jac...@custserv.com writes:
Might be related to
Thanks all for the responses. I'm referring to private below the line.
Unfortunately, there's no dump or map of storage to do any meaningful
debugging. The obvious reason is that something was added somewhere for
private to drop 1M. I'm aware it changes in 1M chunks. However, 'nothing
changed' is
How different was the I/O configuration? Adding devices can reduce private.
Different devices can take different storage amounts for UCB and drivers.
Different hardware can have different recovery modules loaded, changing private
size.
We recently dynamically added DASD. O problem, even over
stars...@mindspring.com (Lizette Koehler) wrote:
So private area below the line is what is left over after the system carves out
everything else.
So CSA, HSA (I think - Hardware area), etc
I was looking for a graphic that displays this better than I can write it.
snip
A reasonable
A pocket protector's worth of Display Commands.
http://idcp.marist.edu/pdfs/ztidbitz/MVS%20CONSOLE%20COMMANDS.pdf
In a message dated 8/27/2015 3:55:31 P.M. Central Daylight Time,
ee...@us.ibm.com writes:
Of course, you want to recheck the end of the specification and the next
1M
On 27 August 2015 at 12:32, Scott Ford idfzos...@gmail.com wrote:
Thanks All, I am going to code up a prototype and test on our Sandbox, so i
can see the results of a 'C' command for example.
BTW, you also mentioned the P (STOP) command. This doesn't lead to
any kind of x22 abend, and is a
If the IODF has more LOCANY=NO devices, you can lose virtual storage below the
line to UCBs.
Bob
On 8/27/2015 3:42 PM, phil yogendran wrote:
Thanks all for the responses. I'm referring to private below the line.
Unfortunately, there's no dump or map of storage to do any meaningful
debugging.
In 55df2edd.5090...@bremultibank.com.pl, on 08/27/2015
at 05:38 PM, R.S. r.skoru...@bremultibank.com.pl said:
What I understod from the lecture:
a) mainframes are old, obsolete, but unfotunately sometimes still in
use - which is a sin.
If they do the job as well as or better than available
All,
Thank you very much for all the informative comments. I am familiar with
the layout of an address space but the refresher is always welcome.
Thanks John, the comments about CSA allocation was particularly
instructive.
As many of you pointed out, I too now believe that changes to the I/O
I did not see in this thread any mention of who owns the DR environment. We own
and manage our own DR environment, so we know exactly what's there. If OP's DR
is in the hands of a vendor, then the I/O configuration may differ as others
have indicated. In particular, UCB location can make a big
On 2015-08-26 12:31, Abe Kornelis wrote:
I have always interpreted as qualified by the qname, though I will
gladly admit this is an inference not directly implied by the text
itself, and therefore subject to discussion.
(not intending to start one right now, though).
This seems amply
Radoslaw Skorupka wrote:
c) IMHO it is bad idea to make any system directly accessible from
Internet. Mainframe, any kind of Unix, Linux, Windows...
Which leaves...what? Is Wang still selling machines? (But those were
systems, too...)
mike.a.sch...@gmail.com (Mike Schwab) writes:
How about Multics? Designed from the start to be multi-user and
highly secure.
some of the CTSS people went to the 5th flr and did Multics. Other of
the CTSS people went to the IBM science center on the 4th flr and did
cp67/cms, the internal
Hi All,
I’m actually the person interviewed in this (frankly overblown) article.
Thankfully I had a chance to talk again about this project here:
https://www.bostonglobe.com/ideas/2015/08/13/remote-corner-internet-art-sprouts/joPVVFqBnctHanbtUBLhzL/story.html
How about Multics? Designed from the start to be multi-user and highly secure.
On Thu, Aug 27, 2015 at 11:19 PM, Timothy Sipples sipp...@sg.ibm.com wrote:
Radoslaw Skorupka wrote:
c) IMHO it is bad idea to make any system directly accessible from
Internet. Mainframe, any kind of Unix, Linux,
New field in V2R1.
Why would IBM make it only 8 characters long? Sigh.
Charles
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf
Of Oren, Yifat
Sent: Thursday, August 27, 2015 1:22 AM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: SMF Type 30
For starters I'd run this job and publish the results here:
Jobcard...
//STEP1 EXEC PGM=IDCAMS,REGION=0M
//SYSPRINT DD SYSOUT=*
//SYSINDD *
LISTC ENT(QWS.SEC.RULES) ALL
/*
I'd also consider changing the DISP=OLD to DISP=SHR as I'm wondering if
something running has the file allocated.
I have reviewed all the replies, but it seemed no progress is made, so here is
my shot at this...
Nathan Astle wrote:
I can view the failing dataset using 3.4 from the driving system.
And also on the ACF system you're trying to IPL? I really hope you have
fallback volsers, VSAM, catalogs, etc
Hi
The SHAREOPTN is (1,3) for VSAM file. Are there a reason for behind the
failure ? Some of the google hits for the error message says improper
catalog sharing.
Nathan
On Thursday 27 August 2015, Nathan Astle tcpipat...@gmail.com wrote:
Hi,
The IKJ message what I am getting is :
Nathan Astle wrote:
The SHAREOPTN is (1,3) for VSAM file. Are there a reason for behind the
failure ? Some of the google hits for the error message says improper catalog
sharing.
I would rather look at ACF recommendation about SHAREOPTN, not at Google.
Groete / Greetings
Elardus Engelbrecht
Does the catalog entry point to the correct volume?
-Original Message-
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
Behalf Of Nathan Astle
Sent: Wednesday, August 26, 2015 9:56 PM
To: IBM-MAIN@LISTSERV.UA.EDU
Subject: Re: Catalog entry not visible
Hi
Thanks Scott for the definitive answer, it does clear up a lot.
Like I mentioned, it's not a game breaker really. I can't come up with a
situation where this behavior would lock us out of our machine. Unless we
really dropped the ball somewhere in the first place. It's more that the list
might
Hi,
The IKJ message what I am getting is :
IKJ56228I DATA SET QWS.SEC.RULES NOT IN CATALOG
OR CATALOG CAN NOT BE ACCESSED
Explanation: DISP=OLD was specified. The dynamic allocation error code is
1708, 5708, or 5710. For a description of the dynamic allocation
return, informational, and error
While we are at it -
SMF30_Highest_Task_CPU_Program has the program name from the task that
reported the largest CPU time percentage for the step (in 30-4, if I understand
the manual correctly).
Regards,
Yifat
-Original Message-
From: IBM Mainframe Discussion List
On Thu, Aug 27, 2015 at 3:22 AM, Nathan Astle tcpipat...@gmail.com wrote:
Hi,
The IKJ message what I am getting is :
IKJ56228I DATA SET QWS.SEC.RULES NOT IN CATALOG
OR CATALOG CAN NOT BE ACCESSED
Explanation: DISP=OLD was specified. The dynamic allocation error code is
1708, 5708, or
73 matches
Mail list logo